Novel Image Analysis-Based Method for Residence Time Distribution Analysis in Steelmaking Tundish

Abstract

A digital image processing-based technique has been developed for residence time distribution (RTD) analysis in a steelmaking tundish. This work presents an alternative approach to current RTD analysis techniques like tracer dispersion analysis, particle image velocimetry, numerical analysis, etc. The video analysis method has been developed, and different process conditions have been evaluated. The physical model of a steelmaking tundish having water as a working fluid similar to the one which is used for RTD analysis through other methods has been used. New incoming fluid (water) has been identified using colored water so that all properties apart from color remain the same. The proposed method breaks the video into static frames, and through application of Lucas–Kanade method, the moving colored pixels in the region of interest are identified. Further intensity analysis of the moving colored pixels is done. In the present novel method, RTD curve is obtained by plotting the intensity of blue component at the outlet vs time. The obtained RTD from VT-LK method has been validated against conventional and established RTD technique widely used for RTD analysis.